Ahmad Mousa

Utilizing cone penetration test (CPT) data to predict shear wave velocity in soil has been widely adopted by engineers over the past few decades. Recently, machine learning (ML) models have emerged as a more efficient tool for predicting shear wave velocity. However, questions remain unanswered on the global applicability, data completeness, and the overall feasibility using of ML models compared to conventional regression methods. In this study, we assemble a global archive linking CPT measurements to shear wave velocity ( V _s ). Prediction strategies are evaluated under grouped, nested cross-validation to emulate deployment to new soundings and sites. The database is curated into five practice-reflective feature regimes and analyzed with decision trees, random forest, gradient boosting, support vector regression (SVR), and a shallow neural network (NN). Across subsets, corrected cone resistance ( q _t ) emerges as the backbone predictor while pore pressure ( u ₂ ) is the highest value complement in clays. Tree ensembles attain the highest test accuracy (low-to-mid 0.8s R ² ) in d + q _t + u ₂ , while SVR and NN are competitive when u ₂ is present. Investigating dataset size shows steep early gains from adding clay-rich records when inputs include u ₂ . The returns, however, diminish beyond a few hundred samples; enhancing q _t -only sets is effective but cannot substitute drainage information. Relative to representative empirical correlations, flexible models improve R ² by roughly 0.04–0.1 for feature sets that capture drainage or stress-history interactions, while regressions remain competitive in the austere q _t -only regime. We conclude through a practical lens: prioritize q _t + u ₂ acquisition where feasible, expand archives when sensors are limited, and apply group-driven validation to secure transferable, bias-controlled estimates of V _s .

The presence of vegetation in urban soil is both visually appealing and environmentally beneficial. Vegetation contributes to changes in soil water content (SWC) and soil suction (SS) through transpiration. These changes ultimately affect slope stability, soil erosion, runoff generation, and water infiltration. This study closely examines the impact of tree canopy covers on SWC and SS dynamics in an urban slope site with a diverse mix of vegetation (trees and grass) in a tropical region. Field monitoring revealed that slope areas with higher tree canopy cover consistently exhibited lower SWC and higher SS at shallower depths, highlighting the stabilizing effects of vegetation. During wet periods, the difference in SWC between high and low canopy cover areas increased by 36% compared to dry periods, emphasizing the role of canopy cover in regulating soil moisture under varying climatic conditions. Furthermore, stations with lower canopy cover demonstrated a stronger correlation between SWC and relative humidity (R values of 0.64 in dry and 0.71 in wet periods) than those with higher canopy cover (R values of 0.57 in dry and 0.63 in wet periods), indicating distinct soil moisture retention patterns based on canopy density. These findings underscore the critical role of tree canopy cover in controlling SWC and SS dynamics, particularly for urban slope stability and sustainable landscape management.

Municipalities and transportation departments worldwide are striving to keep road pavement conditions acceptable, thus enhancing pavement sustainability. Although the pavement condition index (PCI) is widely used to assess distress conditions, traditional visual surveys used for PCI estimation can be laborious, expensive, and time-consuming. The international roughness index (IRI) can be measured more economically and conveniently than PCI; however, it does not directly indicate the surface condition of the pavement. In this study, a PCI–IRI correlation is proposed for urban roads located in the New Beni-Suef region, Egypt. For this purpose, a total of 44 km of urban roads was divided into homogenous sections. A visual distress survey was conducted to measure PCI considering typical distress patterns. The IRI values for the same sections were measured using an ultrasonic distance sensor mounted on an automobile. An exponential model was proposed to capture the relationship between IRI and PCI. With a coefficient of determination of 0.82, the exponential model seems to outperform reported IRI-PCI correlations. Model validation, along with a comparison to the existing models, supports its applicability to a wide range of roads. The proposed model provides a cost-effective means for accurately predicting PCI based on IRI, which is particularly useful for pavement maintenance management programs on limited budgets.